EE-612: Lecture 22: CMOS Process Steps Mark Lundstrom Electrical - - PowerPoint PPT Presentation

Lundstrom EE-612 F06 1

EE-612: Lecture 22: CMOS Process Steps

Mark Lundstrom Electrical and Computer Engineering Purdue University West Lafayette, IN USA Fall 2006

www.nanohub.org

NCN

Lundstrom EE-612 F06 2

• utline

1) Unit Process Operations 2) Process Variations

Lundstrom EE-612 F06 3

unit process operations

1) Oxidation 2) Diffusion 3) Ion Implantation 4) RTA/RTP 5) Chemical Vapor Deposition 6) Lithography 7) Etching 8) Metalization 9) Well Structures 10) Isolation 11) Source / Drain structures

Lundstrom EE-612 F06 4

useful references

1) J.D. Plummer, M.D. Deal, P.B. Griffin, Silicon VLSI Technology, Fundamentals, Practice, and Modeling, Prentice Hall, Upper Saddle River, NJ, 2000. 2) S.A. Campbell, The Science and Engineering of Microelectronic Fabrication, 2nd Ed., Oxford Univ. Press, New York, 2001.

Lundstrom EE-612 F06 5

• xidation

O2 or H2O +carrier gas

fused quartz furnace tube

T = 800 - 1100 oC

heater wafers

Si + O2 → SiO2 Si+ H2O → SiO2 (dry) (wet)

Lundstrom EE-612 F06 6

• xidation and doping

C x

phophorous m > 1

m = CSi CSiO2

boron m < 1

Lundstrom EE-612 F06 7

local oxidation

Si Si3N4 SiO2

Lundstrom EE-612 F06 8

local oxidation (LOCOS)

Si Si3N4 SiO2 'bird's beak' ‘field

• xide’

Lundstrom EE-612 F06 9

constant source diffusion

CS x

dopant-containing gas (e.g. POCl3)

C time C(x) = CS erfc x / 2 Dt

( )

Q = C x,t

( )

∞

∫

dx=2CS Dt π

Lundstrom EE-612 F06 10

Limited source diffusion

CS x

t = 0 ‘predep’

C time C x,t

( )= Q /

πDt

( )exp − x / 2 Dt ( )

2

⎡ ⎣ ⎢ ⎤ ⎦ ⎥

Lundstrom EE-612 F06 11

diffusion

Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si substitutional interstitialcy interstitial “oxidation enhanced diffusion’ D(T ) = D0e−EA /kBT

Lundstrom EE-612 F06 12

ion implantation

• B
• F = Q r

υ × r B magnet ion source acceleration deflection wafer energetic ions bombard silicon wafer I

Lundstrom EE-612 F06 13

ion implantation

Si RP(E) ΔRP(E) implant damage (anneal) N x

( )= N p exp − x − Rp

( )

2 2ΔRp 2

⎡ ⎣ ⎤ ⎦ Q = 2πN pΔRp

Lundstrom EE-612 F06 14

ion implantation (ii)

10 100 1000

P r

• j

e c t e d r a n g e ( μ m ) Acceleration energy (keV)

0.01 0.1 1.0 B As

Lundstrom EE-612 F06 15

channeling

x C x

( )

tilted 3 deg

RP ΔRP

Lundstrom EE-612 F06 16

rapid thermal annealing

lamps reflector quartz window wafer gas inlet

thermal budget Dt

Lundstrom EE-612 F06 17

chemical vapor deposition

wafer

Dt susceptor

gas inlet gas exhaust reaction chamber

2SiH4 + 4NH3 → Si3N4 +12H2 SiH4 → Si+2H2 SiH4+O2 → SiO2+2H2 silicon nitride poly silicon silicon dioxide

Lundstrom EE-612 F06 18

plasma CVD / etching

Dt

gas inlet gas exhaust heater wafer RF power in

lower temperature reduces thermal budget

Lundstrom EE-612 F06 19

lithography

• ptical source

wavelength, λ shutter mask resist wafer

expose, develop, etch

lens contact

• r

proximity

Lundstrom EE-612 F06 20

projection printing

UV source lens 1 lens 2 wafer mask

α

Lundstrom EE-612 F06 21

registration errors

E E

misalignment

E E E E E E E E E

run out

Lundstrom EE-612 F06 22

phase shift lithography

electric field at mask conventional mask phase shift mask intensity at wafer electric field at mask intensity at wafer

Lundstrom EE-612 F06 23

pattern transfer

resist:

• ptically sensitive polymer which,

when exposed to UV changes its solubility in specific chemicals wafer wafer wafer negative resist (less soluble after exposure) positive resist (more soluble after exposure)

Lundstrom EE-612 F06 24

etching

undercut wafer wet chemical etching (isotropic) dry etching (plasma or reactive ion etching - RIE) (anisotropic) wafer chemicals react with underlying material, but not resist ionized gases react with underlying material, but not resist

Lundstrom EE-612 F06 25

pattern transfer (ii)

mask lithography bias chrome resist

LDrawn

etch bias

LGate(physical)

Lundstrom EE-612 F06 26

metalization

www.itrs.net 2005 Edition Tungsten (W) plugs for first layer metal dep CMP

Lundstrom EE-612 F06 27

• utline

1) Unit Process Operations 2) Process Variations

Lundstrom EE-612 F06 28

discrete doping effects

N+ N+ P (NA cm-3)

V = W × L × xj

example: L = 50 nm W = 100 nm xj = 25 nm NA = 1018 cm-3 NTOT = 125 Number of dopants in the critical volume is a statistical quantity

Lundstrom EE-612 F06 29

discrete doping effects (ii)

3D transport leads to inhomogeneous conduction

(see Wong and Taur, IEDM, 1993, p. 705)

Effects: 1) σVT (10’s of mV) 2) lower avg. VT (10’s of mV) 3) asymmetry in ID source drain

Lundstrom EE-612 F06 30

discrete doping effects (iii)

(simulations from A. Asenov group,

• Univ. of Glasgow)

35 nm MOSFET AFM measurements, Fujitsu

Lundstrom EE-612 F06 31

statistical variability

Line edge roughness discrete dopants

From A. Asenov, Univ. of Glasgow

Lundstrom EE-612 F06 32

variability is becoming a major issue

• G. Declerck, Keynote talk, VLSI Technol. Symp. 2005

Lundstrom EE-612 F06 33

• utline

1) Unit Process Operations 2) Process Variations

For a basic, CMOS process flow for an STI (shallow trench isolation process), see: http://www.rit.edu/~lffeee/AdvCmos2003.pdf

Lundstrom EE-612 F06 34

CMOS process flow

For a basic, CMOS process flow for an STI (shallow trench isolation process), see: http://www.rit.edu/~lffeee/AdvCmos2003.pdf The author is indebted to Dr. Lynn Fuller of Rochester Institute of Technology for making these materials

• available. What follows is a condensed version of a

more complete presentation by Dr. Fuller. I regret any errors that I may have introduced by shortening these materials.

• Mark Lundstrom 10/19/06